#exult@irc.freenode.net logs for 21 Jun 2001 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage


[00:06:50] --> chimera|wookin has joined #exult
[01:05:46] <-- matt0 has left IRC (Laserdisc Arcade Emulation Project: http://daphne.rulecity.com , get your CVS password today! =])
[01:38:28] --> matt0 has joined #exult
[01:50:19] <-- matt0 has left IRC (Laserdisc Arcade Emulation Project: http://daphne.rulecity.com , get your CVS password today! =])
[06:17:43] --> Kirben has joined #exult
[06:25:44] <-- Kirben has left IRC (Read error to Kirben[co3007967-a.brasd1.vic.optushome.com.au]: Connection reset by peer)
[07:22:19] --> mrod has joined #exult
[07:23:04] <-- mrod has left #exult
[08:47:11] --> Kirben has joined #exult
[09:40:17] <-- Kirben has left IRC (Read error to Kirben[co3007967-a.brasd1.vic.optushome.com.au]: Connection reset by peer)
[09:40:25] --> Wumpus has joined #exult
[10:11:27] --> Kirben has joined #exult
[10:53:30] --> Nadir has joined #exult
[10:53:45] <Nadir> hi Kirben
[10:53:55] <Kirben> Hi
[11:47:07] <-- Kirben has left IRC (System Meltdown)
[11:51:33] <-- Nadir has left IRC (I feel like dumping a core)
[12:11:07] --> Kirben has joined #exult
[12:41:01] <-- Kirben has left IRC (System Meltdown)
[13:04:53] --> Colourless has joined #Exult
[13:11:55] --> Nadir has joined #exult
[13:12:01] <Nadir> hi colourless
[13:12:10] <Colourless> hi
[13:12:40] <Nadir> Seen my post on memleaks ?
[13:12:55] <Colourless> yeah i noticed it
[13:14:10] <Nadir> memprof is cool :)
[13:14:35] <Colourless> i hear it kind of slows things down
[13:14:38] <Colourless> a lot
[13:15:41] <Nadir> no. That's ElectricFence
[13:15:51] <Colourless> ah
[13:16:21] <Nadir> ElectricFence works by allocating some extra memory above each allocation to catch overwrites.
[13:17:32] <Colourless> ah. MSVC does something similar in debugging mode
[13:34:51] <-- Nadir has left IRC (Read error to Nadir[natmi191.etnoteam.it]: Connection reset by peer)
[13:40:20] --> Nadir has joined #exult
[14:09:00] --> fingolfin has joined #exult
[14:09:03] <fingolfin> hiya
[14:09:09] <Nadir> hi
[14:09:11] <fingolfin> I just got xchat running on OS X ;)
[14:09:11] <Colourless> hi
[14:09:17] <Nadir> good stuff.
[14:09:30] <fingolfin> but I had to disable Perl support in it :/ there are some linking problems..
[14:09:38] <fingolfin> Tristan!!! long long time no see ;)
[14:12:15] <Nadir> :)
[14:12:39] <Nadir> Well you haven't been round here that much either ;)
[14:14:36] * fingolfin takes a screenshot with Gimp, and posts it to his web page
[14:14:44] <fingolfin> tristan: yeah...
[14:15:06] <fingolfin> tristan: I had to rewrite parts of MacOS X´ PPPoE driver to be able to go online again...
[14:15:10] <fingolfin> that took me some time ;)
[14:15:43] <fingolfin> http://fingolfin.dnsalias.net/~maxhorn/xchat.jpg
[14:15:53] <fingolfin> that is X on X ;)
[14:16:44] <Nadir> gimp, x-chat, what else ?
[14:17:01] <-- Colourless has left IRC (Ping timeout for Colourless[ppp556.adelaide.on.net.au])
[14:21:40] <Nadir> Aha WindowMaker !
[14:23:22] <fingolfin> lyx, too
[14:23:24] <fingolfin> gnome
[14:24:48] <Nadir> can you use BSD's port system ?
[14:25:27] --> Colourless has joined #Exult
[14:26:40] --> wjp has joined #exult
[14:26:45] <wjp> hi everyone
[14:27:34] <Colourless> hi
[14:28:50] <Nadir> hi wjp
[14:39:01] <fingolfin> hi wjp
[14:39:06] <fingolfin> Nadir: well, no idea...
[14:39:19] <fingolfin> Nadir: I think this is available for OS X, but I am not using it
[14:39:43] <fingolfin> I rather use "fink" (fink.sf.net) which is based on the debian package system, and which aims to provide packages for many many Unix apps
[14:41:57] * wjp is trying memprof
[14:42:40] --> mrod has joined #exult
[14:42:50] <mrod> hello all, i have found a bad bug
[14:43:10] <mrod> i can reproduce it, and i have a savegame ready to submit
[14:43:33] <Colourless> what's the problem?
[14:44:53] <mrod> briefly: i get the musket+ammo in lb's castle and give them to spark, and save; the musket is still there; but if i reload the game i've just saved, the ammo are not there anymore
[14:45:56] <mrod> i can write more precise instruction and submit the bugreport, but who should i give the savegame to ?
[14:46:25] <wjp> could you try disabling SI paperdolls?
[14:46:26] <Colourless> you can attach files to the bug reports
[14:46:32] <wjp> (F2, P)
[14:46:52] <mrod> Colourless: oh, ok then i will do it, nice
[14:47:18] <mrod> wjp: ok, i'm going to try it after i submit the report
[14:48:04] <wjp> when I drop the musket+ammo on Iolo, the musket appears on his belt, and the ammo in the quiver
[14:48:12] <wjp> in paperdoll mode, both are invisible
[14:48:26] <wjp> well, semi-invisible
[14:48:53] <mrod> i've dropped directly the musket on spark's right hand, and the 50 ammos on his left hand
[14:49:13] <mrod> yes i have paperdoll mode enabled
[14:50:08] <mrod> wjp: wait, you mean that they could still be there ?
[14:50:11] <wjp> yeah
[14:53:51] <mrod> wjp: you're right, they are indeed. i did not think of it, because i dropped them right on spark's hand and expected to find them there, so it didn't come to my mind
[14:54:20] <mrod> they were hidden under spark's armor
[14:54:52] <mrod> but, shouldn't i have found them in the place i've put them ? so i guess there's still a (smaller) bug, isn't it ?
[14:55:09] <wjp> Colourless?
[14:57:06] <wjp> mrod: isn't a musket a two-handed weapon?
[14:57:26] <Colourless> wjp: it's a bug
[14:57:57] <Colourless> things shouldn't move around but they do
[14:58:43] <wjp> and even if they do move around, they shouldn't disappear in the paperdoll
[14:59:02] <Nadir> wjp: what do you think of memprof ?
[14:59:12] <wjp> looks useful
[14:59:26] <Nadir> I like the icon for leaks
[14:59:29] <mrod> wjp: yes it is, but after having put the musket in spark's right hand, the left hand did nevertheless accept the ammos
[14:59:33] <wjp> Nadir: we're losing _thousands_ of nodes...
[14:59:57] <Nadir> literally
[15:00:05] <wjp> mrod: that sounds like another bug :-)
[15:00:10] <Nadir> hi mrod
[15:00:15] <mrod> hello Nadir :-)
[15:00:20] <mrod> wjp: :-)
[15:00:48] <wjp> that could be (one of the reasons) why it's moving them around on loading
[15:02:13] <mrod> hmm, so i guess you don't need the savegame (nor the bugreport) ... what shall i do ?
[15:02:21] <Colourless> wjp: they can disappear becauase i only allow for world shapes to be drawn for lhand and rhand
[15:02:41] <wjp> Colourless: ah, and not for belt, etc...?
[15:02:47] <Colourless> no
[15:03:02] <wjp> Colourless: the funny thing is that it does show the ammo in the quiver
[15:03:32] <wjp> Colourless: it shows a blue line, but it displays the ammo (from shapes.vga) in the quiver itself
[15:04:01] <wjp> Nadir: hehe, yes, the 'leaks' icon is great :-)
[15:05:53] <mrod> ok, i assume you don't need them.
[15:06:24] <wjp> mrod: thanks for the report
[15:06:34] <wjp> mrod: looks like this area needs some more tinkering :-)
[15:07:09] <Colourless> wjp: (about quiver) it shouldn't unless someone changed it
[15:07:12] <mrod> wjp: figure it, i'm so grateful to you all for the whole thing
[15:09:02] <wjp> Colourless: http://www.math.leidenuniv.nl/~wpalenst/paperdolls.png
[15:09:55] <wjp> left paperdoll: hand-overflow, right paperdoll: ammo in quiver
[15:10:22] <Nadir> wjp: where would it be safe to delete the entire Search_node chain ?
[15:11:10] <Colourless> hmm, odd.
[15:11:14] * Colourless checks the code
[15:11:53] <wjp> Nadir: dunno... never looked at the pathfinder code
[15:12:27] <mrod> Nadir: i've downloaded eaw patches
[15:12:38] <mrod> and got sfx
[15:14:50] <Nadir> mrod: excellent. How long did that take ?
[15:15:14] <Nadir> wjp: check out line 317 in path.cc
[15:15:44] <wjp> Nadir: interesting...
[15:15:48] <mrod> Nadir: well, it did took a bit :-)
[15:16:29] <Nadir> wjp: it is doing a lookup.clear() but we obviously need to kill each node at a time...
[15:16:51] <wjp> yeah, I wonder if that code still hangs exult if we enable it now
[15:17:10] * Nadir is trying a recompile
[15:17:22] <Colourless> wjp: ok, from looking at the code, it should always show objects.
[15:17:50] <Colourless> problems may occur if objects get drawn over by other parts of the paperdoll
[15:19:00] <Colourless> wjp: however, I'm probably going to fix the problem at the source... the container game obj loading and saving functions.
[15:19:25] <Nadir> wjp: it hangs
[15:19:37] <wjp> Colourless: the bug with two-handed weapons is separate from that, right?
[15:22:55] <wjp> Nadir: it's going in circles... 0x8575358, 0x85753e8, 0x8575358, etc..
[15:23:57] <Colourless> wjp: perhaps, but it may not be though. I'll have to check it all out.
[15:24:09] <Nadir> why ???
[15:24:27] <wjp> Nadir: maybe hash_set is broken; I'm trying with set now
[15:24:29] <Colourless> because it's broken?
[15:25:17] <Nadir> hash_set is broken ?
[15:25:27] <wjp> it seems to work with a set
[15:27:17] <mrod> i have to disconnect for some time now. see you later
[15:27:22] <wjp> bye
[15:27:26] <-- mrod has left #exult
[15:28:42] <wjp> argh.... these STL header files are practically unreadable
[15:30:19] <fingolfin> wjp: gcc´s ? ;)
[15:30:19] <fingolfin> wjp: but for real, they are unreadable for any implementation I think
[15:30:20] <fingolfin> that´s because of all the template stuff
[15:30:29] <fingolfin> hm, how do I switch on time stamps in x-chat?
[15:30:41] * Wumpus looks at fingolfin... Dave?
[15:30:57] <wjp> settings->setup->Interface->IRC I/O->Output box->Time Stamp all Text
[15:31:26] <wjp> or settings->setup->IRC->Logging->Always timestamp logs to timestamp the logs
[15:32:21] <fingolfin> Wumpus?
[15:32:41] <fingolfin> wjp: thx
[15:35:45] <Nadir> wjp: gcc's STL headers are unmodified
[15:35:58] <Nadir> sorry, that was meant for fingolfin
[15:36:02] <Wumpus> fingolfin: you're not a dave are you?
[15:36:52] * Wumpus just knew (knows, i guess, although I rarely see him now) a `fingolfin' from around here :) and I can't imagine it being that common a nick :)
[15:37:33] <Nadir> Max "Dave" Horn "Fingolfin"
[15:37:41] <Wumpus> hehe
[15:37:42] <Wumpus> okay
[15:37:56] <Wumpus> not the one I know then :) sorry
[15:38:39] * Colourless knows a Dave and a Wumpus from somewhere else, but this wumpus isn't that wumpus
[15:38:45] <Wumpus> hehe
[15:40:04] <fingolfin> lol
[15:40:13] <fingolfin> Max Albrecht Sebastian Horn is my full name, BTW ;)
[15:40:33] <fingolfin> and nadir: what do you mean, gcc´s STL headers are unmodified? unmofied compared to what?
[15:40:40] <Nadir> to SGI's
[15:40:53] <fingolfin> Nadir: ugh, you mean they are that bad??? ;)
[15:41:00] <fingolfin> btw, STL is *not* the standard C++ lib ;)
[15:41:11] <fingolfin> also people keep misusing the word to mean that
[15:43:26] <Nadir> I know that. It's just that gcc until 2.8 had modified (read crippled) STL headers because it couldn't handle them. Only in 2.95 they managed to support the standard SGI headers.
[15:43:48] <Wumpus> oh well
[15:43:51] <Wumpus> bah
[15:43:57] <Wumpus> too many windows and too little sleep
[15:43:58] <Wumpus> :P
[15:44:10] <fingolfin> s/also/although/
[15:44:11] * fingolfin like mixing up words that are pronounced allike... (k)now, (k)night, also/although, you get the idea ;)
[15:44:11] <Nadir> Actually STL is the product of HP (C) 1994 and SGI (C) 1996,1997
[15:44:55] <Nadir> should we use set instead of hash_set then ? is it a better choice anyway ?
[15:45:15] <wjp> Nadir: I think I'm onto something... hang on a sec
[15:46:24] <wjp> ok, I think I fixed it
[15:46:47] <wjp> since the hash-function needs the Search_node, you can't delete it, and then increase the iterator
[15:47:06] <wjp> (the iterator::operator++ calls the hash function)
[15:48:05] <fingolfin> yeah
[15:48:24] <fingolfin> you have to be careful, normally it is completly forbidden to modify a containe while you are iterating it...
[15:48:34] <fingolfin> some containes make exceptions, but in general, don´t od it ;)
[15:48:44] <fingolfin> and hash_set is not standarized
[15:49:06] <fingolfin> (you´ll see that I had to add some extra code for MacOS due to this, as the header files were named differently)
[15:49:07] <Colourless> that is probably why it's fine in win32
[15:49:24] <wjp> yeah, and the using metrowerks::hash_set
[15:49:34] <fingolfin> a large subset of STL is a subset of the standard C++ lib ;)
[15:49:36] <fingolfin> wjp: yeah
[15:49:48] <wjp> fingolfin: do you think it's safe to "delete *(X++)'
[15:49:49] <fingolfin> wjp: they added it to their own namespace, just as the standard *requires* it
[15:50:00] <wjp> fingolfin: :-)
[15:50:23] <fingolfin> wjp: ack! in general, no. in special cases, maybe. depends on which container
[15:50:30] <wjp> hash_set :-)
[15:50:46] <fingolfin> wjp: why not foo *old; old = X; ++X; delete old;
[15:51:03] <Nadir> fingolfin: have a look at the code in path.cc and see what our resident STL-meister can do :)
[15:51:04] <fingolfin> BTW, in general never use X++, but prefer ++X, as it generates better code
[15:51:13] <fingolfin> <g>
[15:51:16] <wjp> fingolfin: that's not the point I meant :-)
[15:51:40] <fingolfin> wjp: sorry... what did you mean?
[15:52:10] <wjp> fingolfin: *X is a pointer to a Search_node. The hash functions still needs this search node
[15:52:31] <fingolfin> BTW, I am trying to use SDL-1.2.1 which has much better OS X support... but the dist is missing some files :/
[15:53:10] --- fingolfin is now known as Fingolfin|dinner
[15:53:23] <Fingolfin|dinner> wjp: you mean the hash function of the hash_set?!?! I am a bit confused
[15:53:30] <Fingolfin|dinner> I ´ll look at it when I am back, OK?
[15:53:41] <wjp> k, but I'll probably be gone then
[15:53:52] <wjp> bye
[15:54:22] <wjp> oh, hmm, I guess I should've mentioned we supply our own hash function :-)
[15:54:28] <Nadir> wjp: does it work with set ?
[15:54:35] <wjp> (which dereferences the pointer stored)
[15:54:39] <wjp> Nadir: sure
[15:55:06] <Nadir> What about memprof results ?
[15:55:16] <wjp> all those 20 byte memleaks are gone
[15:55:36] <Nadir> great !!!
[15:56:33] <Nadir> there are still a few things we don't free, but I guess they are not complex. I really want the ability to go back to the Exult menu when quitting from a BG or SI game
[15:56:52] <wjp> I did have some 8K memleaks from Flex a while ago
[15:57:09] <wjp> Nadir: yeah, that feature would be nice
[15:58:23] <Nadir> memprof provides an easy way to find who allocated those (the entire call chain infact)
[16:00:39] <wjp> I get a 9722 byte leak in Flex.cc, line 97
[16:00:57] <wjp> oh, but now it's gone
[16:01:04] <wjp> (after stopping the music)
[16:01:23] <wjp> and memprof just segfaulted :/
[16:02:34] <Nadir> what memprof are you using ? 0.4.0 or 0.4.1 ?
[16:02:57] <wjp> 0.4.1
[16:04:46] <wjp> I have to go
[16:05:05] <wjp> (I'm committing my changes to path.cc)
[16:05:17] <Nadir> excellent. You're my hero !
[16:05:22] <wjp> :-)
[16:05:35] <wjp> bye
[16:05:40] <Nadir> bye
[16:05:41] <-- wjp has left IRC ([x]chat)
[16:08:31] * Colourless has to go do some stuff in win98
[16:08:34] <Colourless> bbl
[16:08:37] <-- Colourless has left IRC ([1:37])
[16:11:11] <Nadir> I'm going too. Bye all
[16:11:13] <-- Nadir has left IRC (I feel like dumping a core)
[16:39:39] --> Nadir has joined #exult
[16:51:51] <-- Nadir has left IRC (ChatZilla 0.8.2 [Mozilla rv:0.9.1+/20010620])
[16:57:54] <-- Wumpus has left IRC (raise( SIGTSTP );)
[17:26:29] --> Colourless has joined #Exult
[17:27:01] * Colourless is now back
[17:36:13] <Fingolfin|dinner> wb
[17:36:17] --- Fingolfin|dinner is now known as Fingolfin
[17:36:58] <Colourless> wb
[18:08:46] --> InuYasha has joined #exult
[18:13:33] <InuYasha> Moo.
[18:13:50] * Colourless raises an eyebrow
[18:26:03] <-- InuYasha has left IRC (Ping timeout for InuYasha[cpe-24-221-255-164.ut.sprintbbd.net])
[18:33:02] * Colourless is off to bed... before dawn even
[18:33:08] <-- Colourless has left IRC ([4:02])
[18:47:38] <-- Fingolfin has left #exult
[20:26:31] --> mrod has joined #exult
[20:26:55] <-- mrod has left #exult